| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, University of Vienna, Waehringer Guertel 18-20,
A-1090 Vienna, Austria.
2 Department of Internal Medicine, Krankenhaus der Barmherzigen Schwestern,
A-1060 Vienna, Austria.
Received March 28, 2002;
accepted after revision September 10, 2002.
Address correspondence to F. M. Lomoschitz.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. In 113 patients (73 women, 59 ± 14 years old; 40 men, 48 ± 16 years old), dual X-ray absorptiometry of the lumbar spine and the proximal femur, imaging-guided quantitative sonography, and non–imaging-guided quantitative sonography of the calcaneus were performed. The percentage of patients having a T-score equal to or less than a threshold of –2.5 SDs (prevalence of osteoporosis) was calculated for each imaging technique. The diagnostic agreement of the three techniques in identifying individuals with osteoporosis was assessed.
RESULTS. Eleven percent of the women and 8% of the men were classified as osteoporotic by imaging-guided quantitative sonography, and 38% of the women and 25% of the men were so classified by non–imaging-guided quantitative sonography. At dual X-ray absorptiometry of the spine, 44% of the women and 38% of the men were classified as osteoporotic, and, at different femoral regions, 19–60% of the women and 8–38% of the men were so classified. Kappa analysis for both quantitative sonography techniques was not significant. Kappa analysis for both quantitative sonography techniques and dual X-ray absorptiometry showed diagnostic agreement to be generally poor.
CONCLUSION. No advantage in diagnostic accuracy could be found for imaging-guided quantitative sonography. The considerable diagnostic disagreement between both quantitative sonography techniques and dual X-ray absorptiometry could be confusing in daily clinical practice.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Osteoporosis, defined by the World Health Organization for white women who do not have a fracture, is a T-score of more than 2.5 SDs below the peak bone mass in dual X-ray absorptiometry of the spine, hip, or forearm [5]. Men were not included in these recommendations, and the appropriate T-score for the presence of osteoporosis in men has not yet been determined. No specific recommendations exist about quantitative sonography measurements.
The most commonly used technique to examine skeletal status is a projectional measurement by dual X-ray absorptiometry, typically performed at the lumbar spine and the proximal femur. The only established technique for noninvasive assessment of bone status that does not require radiation is quantitative sonography, which measures ultrasound velocity (speed of sound) and broadband ultrasound attenuation, predominantly at the calcaneus. In addition, quantitative sonography at the calcaneus has shown appropriate results in clinical studies and is becoming increasingly available [2, 6, 7, 8, 9]. The inhomogeneity of the calcaneus as a weight-bearing, predominantly cancellous bone has led to the recent development of imaging-guided quantitative sonography scanners that calculate a lateral quantitative sonography image at the calcaneus for improved identification of appropriate measurement areas [10]. However, significant discordance can be found among measurements obtained by different techniques and among measurements at different skeletal sites [8, 10, 11, 12].
The aim of our study was to evaluate the diagnostic agreement between imaging-guided quantitative sonography and non–imaging-guided quantitative sonography of the calcaneus as well as dual X-ray absorptiometry of the spine and the femur.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Examination Techniques
Dual X-ray absorptiometry was performed in all patients with a QDR-4500
scanner (Hologic, Waltham, MA) using the manufacturer's recommended standard
procedures for the posteroanterior lumbar spine at L1–L4 and for the
proximal femur at the femoral neck, trochanter, intertro-chanteric region, and
total region of proximal femur. The projectional bone mineral density values
were given in grams per centimeter squared, and the individual results were
expressed as a T-score. Fractured vertebral bodies, if present, were excluded
from the analysis. Fractures were determined by semiquantitative assessment of
each vertebral body on lateral and anteroposterior conventional radiographs of
the lumbar spine. The presence of fracture was defined as altered morphology
and a decrease in vertebral height of approximately 25% or more at the
anterior, medial, or posterior aspect of the vertebral body
[2,
11]. Conventional radiographs
of the lumbar spine were also evaluated for degenerative changes. Patients
with obvious calcifications that might lead to an overestimation of bone
mineral density were excluded from the study
[13,
14].
At the same session, the patients were examined at the calcaneus with a conventional non–imaging-guided quantitative sonography scanner (Achilles; Lunar, Madison, WI) and an imaging-guided quantitative sonography scanner (UBIS 3000; Diagnostic Medical Systems, Montpellier, France). Both systems were used according to the manufacturer's recommended standard procedures. Both quantitative sonography systems are of the transmission type, with two ultrasound transducers (transmitter and receiver) positioned on each side of the heel.
The quantitative sonography measurement of the stiffness of the calcaneus was obtained with the non–imaging-guided unit. The sonography-derived stiffness was automatically determined by the scanner software according to the following formula: stiffness = (0.67 x broadband ultrasound attenuation) + (0.28 x speed of sound) –420 [6, 15]. Stiffness is the default parameter used by the manufacturer for demographic comparison of the patient's data; thus a T-score is derived.
The quantitative sonography measurement of the broadband ultrasound attenuation and speed of sound of the calcaneus were obtained with the imaging-guided unit. The results of imaging-guided quantitative sonography are represented as a gray-scale image of the calcaneus in a lateral projection (Fig. 1). With this approach, low quantitative sonography attenuation is depicted by dark gray-scale pixels, and high attenuation, by bright gray-scale pixels. Both broadband ultrasound attenuation and ultrasound velocity as absolute values were automatically calculated from a circular region of interest with the lowest quantitative sonography result. This region was automatically identified by the software through a default process and represented the area of lowest attenuation. Quantitative sonography measurement of broadband ultrasound attenuation was expressed as a T-score.
|
For all methods—quantitative sonography and dual X-ray absorptiometry—the diagnostic bone mass threshold for defining osteoporosis in individuals without fractures was based on the World Health Organization [5] criterion of a T-score of less than –2.5 SDs [4].
Statistical Analysis
Means and SDs for all examined parameters were calculated separately for
men and for women. The percentage of subjects with T-scores of less than the
threshold –2.5 SDs was calculated for both groups, for each technique,
and for each parameter. Comparisons of all measured absolute parameters and
T-scores were obtained by pooling all women and all men and performing linear
regression analysis using Pearson's correlation coefficient (r). In
addition, the percentage of standard errors of the estimate (which equals the
coefficient of variation) and p values to test the significance of
correlation were obtained. To compare the diagnostic ability of the
techniques, we performed a kappa score analysis separately for women and men.
This analysis was performed by classifying every subject as osteoporotic if
the T-score with respect to the reference group was less than –2.5
SDs.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Linear regression analysis for comparing absolute values and T-scores showed correlation coefficients in the women that were similar to those found in men (Tables 2 and 3). The absolute measured parameters obtained by both quantitative sonography units correlated reasonably with each other (r = 0.62 in women and r = 0.80 in men) but correlated modestly to poorly with all other measurements, with correlation coefficients of less than 0.58 in women and less than 0.57 in men (Table 2).
|
|
Correlation of T-scores revealed a similar result (r = 0.63 in women and r = 0.80 in men) when both quantitative sonography methods were compared and provided poor results compared with dual X-ray absorptiometry measurements (r < 0.49 in women and r < 0.61 in men) (Table 3). When all techniques were compared, the strongest correlation of T-scores was in one technique, that for dual X-ray absorptiometry of the total region and of the trochanteric region (r = 0.91) (Table 3).
In the women, non–imaging-guided and imaging-guided quantitative sonography correlated moderately (r < 0.65) with each other, and both quantitative sonography techniques showed poor correlation with dual X-ray absorptiometry of the spine (r < 0.45) (Figs. 2, 3, 4).
|
|
|
Kappa analysis of T-scores did not show significant results between both
quantitative sonography techniques (Table
4). Between imaging-guided quantitative sonography and dual X-ray
absorptiometry, the diagnostic agreement was only moderate compared with dual
X-ray absorptiometry measurements at the intertrochanteric region (
=
0.32 in women and = 0.53 in men) and the total region ( = 0.38
in women and = 0.45 in men). Compared with the other regions measured
by dual X-ray absorptiometry, the diagnostic agreement was not significant.
Between non–imaging-guided quantitative sonography and dual X-ray
absorptiometry, the diagnostic agreement was not significant for dual X-ray
absorptiometry measurements at the spine and the trochanter in women.
Comparing non–imaging-guided quantitative sonography with the other
regions measured by dual X-ray absorptiometry, the diagnostic agreement was
generally poor ( = 0.22–0.39)
(Table 4).
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A number of previous studies [4, 6, 7, 8, 9, 10, 11, 12, 16, 17, 18, 19, 20, 21, 22] have shown that all four techniques used in our study (i.e., imaging-guided quantitative sonography of the calcaneus, non–imaging-guided quantitative sonography of the calcaneus, dual X-ray absorptiometry of the spine, and dual X-ray absorptiometry of the femur) have the ability to differentiate between healthy premenopausal and healthy postmenopausal subjects and to reflect age- and menopausal-related bone loss.
Dual X-ray absorptiometry is the most widespread diagnostic modality and is generally accepted as a reliable approach for use in the diagnosis of osteoporosis [1, 2, 11]. Measurements of the spine [11, 23] and the hip [10, 24] have also proven useful in the diagnosis of local and systemic osteoporotic changes. In recent years, quantitative sonography measurements have shown results comparable to those of dual X-ray absorptiometry for the prediction of osteoporotic fractures [7, 16, 17, 25, 26]. Quantitative sonography also has been successfully proposed as a screening tool for osteoporosis [9]. The attractiveness of quantitative sonography, which was introduced at the calcaneus in the late 1980s, lies in its low cost, ease of use, and freedom from ionizing radiation. The structural heterogeneity of the calcaneus has led to the development of quantitative sonography to permit optimal selection of regions of interest [18, 19]. Laugier et al. [18] showed that quantitative sonography improved a reproducible selection of a region of interest at the calcaneus. Fournier et al. [20] found that automatic region-of-interest placement reduced precision errors caused by improved intra- and interobserver variation. With the imaging-guided quantitative sonography scanner used in this study, the region of interest is usually automatically depicted at the zone of the lowest attenuation. This region corresponds to the area that enables fastest traverse of ultrasound waves and therefore represents the area of best reproducibility.
However, one of the main problems with quantitative sonography is the lack of standardization of different scanners [8, 10], as has been achieved for dual X-ray absorptiometry [27]. Measurements of the same patients with sonographic equipment from different manufacturers have yielded significantly different results [11]. With the advent of an additional quantitative sonography technique that enables imaging-guided measurements, it seems necessary to compare both types of scanners with measurements from the same patients.
When dual X-ray absorptiometry and quantitative sonography measurements were directly compared, several investigators [8, 21, 28, 29, 30] found modest correlation at axial and peripheral measurement sites. The correlation between dual X-ray absorptiometry–based methods and sonographic measurements was only marginally lower than those between dual X-ray absorptiometry–based methods at different skeletal sites. Typically, patients were determined to be at risk using different methods with widely varying classifications [8, 11]. These levels of diagnostic disagreement may result from a number of factors. All these techniques are subject to inherent specific error sources as well as basically different methodologies and units of expression that undermine their respective abilities to define true biologic relationships in or among anatomic sites or among individuals [2, 8, 11]. However, even aside from technical considerations, it is likely that genetic and environmental influences and pathophysiologic forces may influence the cortical and trabecular bone envelopes or the appendicular and axial skeleton in a different manner [31], and thus some of the measured discrepancies reflect true anatomic variations. Recent studies [9] have proven that for quantitative sonography, the general results and trends for men and women are comparable, and that cutoff levels for further screening should be defined.
Comparing the two quantitative sonography methods, correlation in men was slightly greater than in women for both absolute values (r = 0.80 in men vs r = 0.62 in women) and T-scores (r = 0.80 in men vs r = 0.63 in women), but without a statistically significant difference. Similar to results reported by Krestan et al. [10], T-scores and absolute values correlated at nearly the same levels in our study, which might be an indication that the reference database supplied by these systems provides reliable results. For T-scores, measurements of both non–imaging-guided and imaging-guided quantitative sonography correlated poorly with dual X-ray measurements. In general, correlation among both sonographic methods and dual X-ray absorptiometry–based methods was poor. However, results may differ using other calcaneal quantitative sonography systems.
Consistent with the low correlation coefficient was the poor diagnostic agreement (by kappa analysis) among these measurements in classifying patients as osteoporotic. In most instances in which a patient was classified differently by two methods, the differences in T-scores were substantial. Considerable differences were also noted in the percentage of the population with scores less than or greater than the given threshold of –2.5 SDs.
As reported by Grampp et al. [9], one factor that influenced our results is the use of different normative databases by the equipment manufacturers. This fact has also recently been shown [22] to account for significant differences in the diagnosis of osteoporosis; however, because of different techniques and a lack of standardization, these differences are unavoidable.
In summary, our study shows that quantitative sonography, either imaging-guided or non–imaging-guided, provides a comparatively weak agreement with dual X-ray absorptiometry of the spine and femur, which is generally accepted as a reliable method for diagnosing osteoporosis. Results derived from the various techniques were only moderately correlated, which, in a substantial number of patients, precluded the ability to define one parameter as more influential than others in the diagnosis of osteoporosis and may certainly cause considerable disagreement in diagnostic classification. No advantages in diagnostic accuracy could be found for the imaging-guided quantitative sonography system compared with the non–imaging-guided quantitative sonography system. The use of either quantitative sonography technique as the exclusive diagnostic method might therefore not be recommended.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  C. R. Krestan, S. Grampp, C. Henk, P. Peloschek, and H. Imhof Limited Diagnostic Agreement of Quantitative Sonography of the Radius and Phalanges with Dual-Energy X-Ray Absorptiometry of the Spine, Femur, and Radius for Diagnosis of Osteoporosis Am. J. Roentgenol., September 1, 2004; 183(3): 639 - 644. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
