| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Bialystok Medical Academy, Sklodowskiej-Curie 24A,
15-279 Bialystok, Poland.
2 Present address: Department of Radiology, Division of Neuroradiology, Hospital
of the University of Pennsylvania, 3600 Market St., Science Bldg., Ste. 370,
Philadelphia, PA 19104.
3 Department of Neurosurgery, Bialystok Medical Academy, 15-279 Bialystok,
Poland.
Received December 27, 2001;
accepted after revision December 31, 2002.
Supported by the Polish State Committee for Scientific Research grant 6
P05C 050 20.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. We prospectively studied 214 consecutive patients (110 male, 104 female; age range, 12–77 years) who were routinely referred for cerebral angiography. Middle cerebral artery spasm was graded as mild (≤ 25% of vessel caliber reduction) and moderate to severe (> 25% reduction). Angle-corrected blood velocity measurements were obtained using a 2.5-MHz probe. The velocity values were reexpressed as a percentage of the mean of normal reference values for the relevant age, for subjects younger than 40 years, and for sex.
RESULTS. The prevalence of spasm among 335 arteries studied was 8.1% for mild and 12.8% for moderate to severe middle cerebral artery narrowing. For distinguishing all or moderate to severe vasospasm from lesser grades of vasospasm, peak systolic velocity was the best parameter. Areas under ROC curves for all and moderate to severe middle cerebral artery spasms were 0.83 and 0.92, respectively. After standardization, the ROC areas increased significantly (p < 0.05) for all, to 0.86, and only slightly, to 0.93, for moderate to severe spasms. For all grades of middle cerebral artery spasm, the best efficiencies were found at standardized velocity value of 170%.
CONCLUSION. The accuracy of transcranial color Doppler sonography is high in the identification of middle cerebral artery spasm. Standardization of velocities with respect to age and sex increases the accuracy of the method in diagnosing mild middle cerebral artery spasms.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Advances in sonographic technology have allowed the use of the transcranial Doppler technique for detecting vasospasm and monitoring its evolution, because blood flow velocity increases with vessel narrowing [5]. Despite numerous published reports about this subject, the problem of accurate sonographic diagnosis of cerebral vasospasm is still far from being resolved [6, 7]. One reason for this relates to shortcomings in conventional transcranial Doppler sonography [8, 9]. A major limitation of this method is the inability of the operator to visualize the intracranial vessel being examined and hence to define the angle between the vessel and the ultrasound beam. Variability in the angle degrades the reproducibility of blood velocity measurements. Moreover, many authors have focused on the methodologic assumption that a specific and universal velocity threshold exists that reliably discriminates cerebral vasospasm from normal vessel status [1, 5, 10–13]. Apart from the notion that vasospasm is not an all-or-none phenomenon, different cutoffs may be more suitable to screening for and definitive diagnosis of vasospasm [6, 7, 14, 15].
The problem of sonographic diagnosis of cerebral vasospasm should be revisited for several reasons. First is the advance in technology, mainly through the introduction of transcranial color Doppler sonography. This method allows visualization of the vessel examined and obtaining angle-corrected blood velocity, which improve reliability and reproducibility of intracranial blood flow velocity measurements [16]. Second, the overall accuracy of both conventional and color transcranial Doppler sonography in diagnosing middle cerebral artery spasm has not been evaluated using receiver operating characteristic (ROC) curve analysis. Third, the proponents of using particular threshold velocities to diagnose cerebral vasospasm have failed to take into account the variation in Doppler blood flow parameters with sex and age [5–7, 10–12, 17]. Because these differences appear both in the healthy population and in patients after subarachnoid hemorrhage [18, 19], the obtained velocity values should be standardized with respect to age and sex before any subsequent analysis. In this article, we give the results of ROC analysis for the performance of transcranial color Doppler sonography in the diagnosis of cerebral vasospasm, based on age- and sex- standardized blood flow velocities.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Two hundred fourteen consecutive patients routinely referred by neurosurgeons for intraarterial digital angiography were studied prospectively. The patients were 110 males and 104 females ranging in age from 12 to 77 years (median, 49 years). The distribution of patients with respect to age is shown in Figure 1. One hundred twenty-eight patients had suffered from intracerebral hemorrhage, including nontraumatic subarachnoid hemorrhage; 64 patients were admitted with suspicion of a cerebral aneurysm with no evidence of intracerebral hemorrhage; eight were diagnosed as having a brain tumor; 10, ischemic stroke; one, Marfan syndrome; and three, traumatic subarachnoid hemorrhage.
|
The patients fit into the following clinical grades according to the modified scale of the World Federation of Neurological Surgeons [20] and on the basis of an estimation of consciousness using the Glasgow Coma Scale (GCS): 95 patients were in grade 1 (GCS score, 15), 82 patients in grade 2 (GCS score, 12–14), 21 in grade 3 (GCS score, 9–11), 12 in grade 4 (GCS score, 6–8), and four in grade 5 (GCS score, 3–5).
Transcranial Color Doppler Sonography
Intracranial cerebral arteries were studied using a sonographic scanner
(SSA 140 A, Toshiba Medical Systems, Tokyo, Japan) equipped with a 2.5-MHz
90° phased array probe for both B-mode and Doppler imaging. The middle
cerebral arteries were studied via temporal acoustic windows using methods
previously described [21]. To
avoid the risk of sampling a branch of the artery, we insonated the horizontal
segment by placing a 3-mm-wide sample volume 10 mm from the carotid
bifurcation. To determine the angle of insonation, we placed a linear marker
provided by the scanner software, under visual guidance, on the color Doppler
image of the arterial segment being insonated, and its direction was fitted to
be oriented along the long axis of the segment. The angle between this marker
and the ultrasound beam, displayed automatically on the screen of the scanner,
was considered a two-dimensional approximation of the angle of insonation.
This procedure allowed the angle-corrected blood flow velocities to be
measured (Figs. 2A,
2B). The mean, peak systolic,
and end diastolic velocities were calculated by tracing the maximum frequency
envelope of the Doppler waveform.
|
|
Angiographic Studies
Selective intraarterial digital subtraction angiography was performed via
the femoral artery using the Seldinger approach in both internal carotid
arteries and in at least one vertebral artery for every patient. Standard
digital subtraction angiographic images included anteroposterior, lateral, and
one oblique view that were obtained routinely at contrast injection rates of 6
mL/sec and filming rates of three frames per second (ARGOS 2M, Mecall, Milan,
Italy). The field of view was 30 cm for all images. Two neuroradiologists who
were unaware of the sonographic findings reviewed all angiograms to detect and
quantify the cerebral vasospasms.
The image showing the most severe middle cerebral artery narrowing was used for comparison with transcranial color Doppler sonographic findings. All measurements were performed using calipers on the digital display. The resolution of this technique is 0.1 mm. To quantify the degree of vasospasm, we measured the horizontal segment of the middle cerebral artery at its point of maximal reduction and compared that with the normal section of the artery adjacent to the narrowed segment (Figs. 2A, 2B). Angiographic vasospasm was graded as none, mild (≤ 25% of vessel caliber reduction) and moderate to severe (> 25% of vessel caliber reduction). Whenever possible, the diameter of the vessel in question was also compared with that of the contralateral artery to facilitate classification. Discordant readings by the observers were resolved by consensus. Interobserver agreement as to detection and quantification of middle cerebral artery spasm was evaluated using the kappa test. Separate ROC analyses were performed for a subgroup of moderate to severe vasospasm and for all grades of vasospasm (i.e., mild plus moderate to severe).
Statistical Analyses
ROC methodology is an effective tool for determining overall accuracy
(performance) and efficiency of a diagnostic test. Accuracy is the ability of
a test to discriminate between health and disease over the complete spectrum
of operating conditions [22].
Efficiency is the percentage (0–100%) of correct classifications of all
classifications at a given position of discriminator
[22].
ROC analysis.—The entire span for each velocity (peak systolic, mean, and end diastolic), found empirically in all the arteries examined, was split into 40 bins to reduce the possible error when comparing areas under the curve for particular velocities [22]. The ROC curve was constructed by graphing the sensitivity on the ordinate as a function of the false-positive rate (or 1 — specificity) for all 40 cutoff values of flow velocity. The sensitivity for each of the 40 velocity thresholds was calculated as the proportion of positive sonographic results among arteries with angiographic vasospasm. Similarly, specificity was computed as the proportion of negative sonographic results among arteries with no angiographic vasospasm.
The performance of transcranial color Doppler sonography in diagnosing cerebral vasospasm was ranked for each velocity by calculating the areas under the curves and, more specifically, the areas covering high-sensitivity regions [23]. The statistical significance of the difference among the particular areas under the ROC curves was calculated using the paired Student's t test with the correction for correlated data because the distribution of observed variables in the nondiseased and diseased arteries could be approximated to a binormal model [22].
Once the parameter for best performance was selected, it was possible to identify a velocity threshold that can optimally distinguish between spasm and nonspasm states. A value was chosen for which the best trade-off between specificity and sensitivity was achieved.
Standardization procedure.—All obtained values of blood flow velocity were reexpressed as a percentage of the mean of normal reference values for a given age range and sex. We based the procedure on the velocity reference data published in the American Journal of Roentgenology [18]. For example, the standardization of an obtained peak systolic velocity of 182 cm/sec would produce a value of 198% for a patient older than 60 years, because the mean normal reference value for this age group has been established at 92 cm/sec (standardized velocity = 182 / 92 x 100% = 198%). The same velocity of 182 cm/sec would yield 167% of the standardized value for patients in the age span 41–60 years because the mean value of peak systolic velocity in this group is 109 cm/sec. For the age group 20–40 years, statistically significant differences between the sexes has been found to exist, so two mean normal reference values need to be taken into account: 125 cm/sec for women and 113 cm/sec for men. Consequently, recalculated standardized values for the velocity of 182 cm/sec would amount to 146% and 161% for young women and men, respectively.
The reference velocity values determined from reference 18 and the present data were both obtained with the same method of color Doppler sonography with adjustment of flow velocity for the angle of insonation. Clearly, each laboratory could use its own normal reference data for this kind of standardization, especially when dealing with conventional transcranial Doppler sonography, for which the reference values given in reference 18 are not valid.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Mild vasospasm was angiographically diagnosed in 27 arteries, and moderate
to severe vasospasm in 43 arteries. The kappa test for interobserver agreement
in angiographic diagnosing of middle cerebral artery spasm resulted in a kappa
value of 0.80. The agreement between two radiologists, who graded middle
cerebral artery spasms as "mild" or "moderate to
severe," appeared to be somewhat lower, at 
= 0.63.
The ROC curves for moderate to severe and all grades of vasospasm are shown in Figures 3 and 4, respectively. In each figure, the ROC curves are based on standardized and nonstandardized values of the peak systolic velocity. The areas under the ROC curves (Az) for the peak systolic, mean, and end diastolic velocities for moderate to severe and for all grades of vasospasm are given in Tables 1 and 2, respectively. Each table compares total (Az total) and partial (Az partial) areas for nonstandardized and standardized blood flow velocities in the middle cerebral artery.
|
|
|
|
From Figures 3 and 4 and the tables, it can be seen that moderate to severe vasospasm can be diagnosed with high accuracy using transcranial color Doppler sonography. The performance of this method for the diagnosis of all grades of vasospasm was somewhat reduced, as illustrated by the 10% lower values of Az. Both total and partial areas under the ROC curves were either higher or equal for the nonstandardized peak systolic velocity than for the end diastolic and mean velocities, respectively, although the difference did not reach statistical significance (Table 1).
Standardization with respect to age and sex appears to slightly improve the performance of all Doppler waveform parameters in diagnosing moderate to severe middle cerebral artery spasm, as expressed by a greater total area under the related ROC curves (Table 1 and Fig. 3). The effect of standardization becomes apparent in the group containing all grades of vasospasm, as seen in Table 2 and Figure 4. For all velocities, the differences reached statistical significance as calculated using the paired Student's t test (p < 0.05). The diagnostic performance of peak systolic velocity after standardization also remained the highest (Tables 1 and 2).
In addition to the overall performance of transcranial color Doppler sonography in diagnosing middle cerebral artery spasm with the ROC analysis, one can determine the diagnostic efficiency of the test for any required blood velocity threshold (e.g., the best trade-off between sensitivity and specificity). Figures 3 and 4 depict superimposed ROC curves for the standardized and nonstandardized peak systolic velocities. The arrows indicate velocity thresholds corresponding to the best efficiency in diagnosing moderate to severe (Fig. 3) and all (Fig. 4) middle cerebral artery spasms. Some measures of test accuracy—such as efficiency, sensitivity, specificity, positive and negative predictive values—linked to these velocity thresholds are given in Tables 3 and 4. The lower threshold yields higher sensitivity with lower specificity, whereas the higher threshold results in lower sensitivity and higher specificity. Although thresholds like these can be useful in screening for vasospasm or for definite confirmation of this condition, usually the efficiency of these thresholds is not optimal. For this reason, a working velocity threshold is often sought, one related to maximum efficiency. In our study, the optimal efficiency for the peak systolic velocity corresponded to 182 cm/sec for detection of moderate to severe middle cerebral artery spasm when nonstandardized values were analyzed. For the standardized values of the peak systolic velocity, the best efficiency corresponded to 170% of the mean normal reference values for a given age span and sex (see the Subjects and Methods section). The optimal efficiency of transcranial color Doppler sonography in the detection of all middle cerebral artery spasms corresponded to 160 cm/sec of the peak systolic velocity and to 170% of a mean normal reference value for standardized velocities (Tables 3 and 4).
|
|
Those who prefer to deal with absolute values rather than with percentages can readily recalculate the obtained threshold of 170% with respect to the mean normal reference values in different age groups [18]. Thus, in patients older than 60 years, the threshold of best efficiency is 156 cm/sec; in patients of 41–60 years, 185 cm/sec; in young women, 213 cm/sec; and in young men, 192 cm/sec.
In the group of moderate to severe vasospasm, efficiency at the optimal blood flow velocity threshold for diagnosing vasospasm was similar for standardized and nonstandardized peak systolic velocity (Tables 3 and 4). However, standardization improved the sensitivity by 6% at the expense of only a 1% decrease in specificity. On the other hand, standardization was found to enhance efficiency when mild vessel narrowing was of interest, such as in the group of all grades of vasospasm. For the optimal threshold of the peak systolic velocity, efficiency increased by 3%; and although sensitivity decreased by 4%, the specificity increased by 5% when compared with nonstandardized values. Standardization increases the positive predictive value of the optimal threshold of the peak systolic velocity by as much as 18% (Tables 3 and 4).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Our results show that standardization of flow velocities with respect to age and sex can further increase the diagnostic performance of transcranial color Doppler sonography in the detection of middle cerebral artery spasm. Indeed, the overall performance of this method for the diagnosis of advanced middle cerebral artery spasm (expressed as the area under the ROC curve) increased only slightly. This result could have been anticipated because significant narrowing of the artery causes an increase in the flow velocity well above the normal range. Consequently, the diagnostic performance of transcranial color Doppler sonography is very high, even for the nonstandardized velocities. Nevertheless, at the best efficiency threshold, sensitivity after standardization increased by 6%.
The effect of standardization becomes statistically significant when mild middle cerebral artery spasms are also included in the studied group, which is expressed by a statistically significant increase of the area under the ROC curve in the group of all vasospasms (Fig. 4 and Tables 2 and 4). The efficiency of the optimal threshold also improved. This substantial improvement in the performance of transcranial color Doppler sonography for diagnosing mild middle cerebral artery spasm is of clinical importance because it allows early diagnosis of the initial phase of the spasm. Early diagnosis can influence neurosurgical planning and may help to more accurately assess the patient's risk of complications from vasospasm at an early stage of the disease process. Standardization reduces the number of false-positive results in young patients and false-negative results in older patients.
It would be interesting to compare the performance of transcranial color Doppler sonography with that of conventional Doppler sonography in diagnosing middle cerebral artery spasm. However, we found comparison difficult because the application of transcranial Doppler sonography has never been studied with ROC methodology. In most studies [5–7, 10–14, 26, 27], only one (or, more rarely, two) velocity thresholds are assessed along with relevant measures of efficiency such as sensitivity, specificity, and positive and negative predictive values. The wide range of blood flow velocity thresholds (100–200 cm/sec mean velocity) proposed by different authors to diagnose middle cerebral artery spasm would seem to bring the usefulness of transcranial Doppler sonography into question.
More detailed critical reviews of published reports have been provided by Lysakowski et al. [28] and Mariak et al. [29]. In those studies, not only proposed diagnostic blood flow velocity thresholds but also measures of efficiency of transcranial Doppler sonography in the diagnosis of middle cerebral artery spasm vary significantly. This variance can be explained in part by differences in the population of patients studied, but above all can be explained by the different prevalence of vasospasm (which can heavily influence the positive and negative predictive values) [22]. Selection of an optimal blood velocity threshold that is diagnostic for middle cerebral artery spasm for a given population can also depend on what kind of intracranial disorders prevail in the population. Thus, with a high prevalence of intracranial hypertension, the diagnostic threshold of blood flow velocity will be lower [15, 29]. On the other hand, conditions such as intracranial hematoma and advanced atherosclerosis can significantly change the course of the major cerebral arteries [9]. As a result, the artery can deviate from the line of the ultrasound beam, which leads to a superficially low blood flow velocity if "blind" transcranial Doppler sonography is used and no angle correction is performed. This source of errors can be eliminated to a great extent with transcranial color Doppler sonography, because this method enables the operator to measure the angle between the vessel and the ultrasound beam [21].
Another source of excessive variability in the results hitherto reported is related to such confounding factors as age and sex [14, 18, 19, 30]. These factors have not been taken into account in most studies of cerebral vasospasm, despite its being well documented that blood flow velocities change significantly with age and sex, both in healthy populations and in patients with disorders, including subarachnoid hemorrhage [18, 19].
By using the color Doppler technique we were able to reduce the error in velocity measurement of the variable course of the middle cerebral artery, and by standardizing the obtained velocities we could decrease the variability introduced by the factors of age and sex. This statement can be further supported by a direct comparison of our results with results obtained with conventional transcranial Doppler sonography by Sloan et al. [14]. Admittedly no ROC analysis was done in their study, but the sample size and prevalence of vasospasm were similar to ours. At the optimal threshold of flow velocity, our sensitivity and specificity were 79% and 96%, respectively, whereas in the study of Sloan et al. these figures were 52% and 79%. Standardization increased our sensitivity to 85% with only a minimal (1%) decrease in specificity.
Despite its many advantages, sonography of cerebral vasospasm is not free from limitations. One most apparent is the inability to detect the echo signal because of the inadequacy of the temporal window. In our study this occurred in 24 patients (11.2%), a proportion similar to that reported with conventional transcranial Doppler sonography [14]. This figure could be decreased by administration of a sonographic contrast agent, which was not used in our study.
Where the examined vessel is not patent, no Doppler signal is to be detected. With conventional transcranial Doppler sonography, this lack of signal can easily be mistaken for the absence of a temporal window. With color Doppler sonography, such occluded arteries can be identified. In our patients with an adequate acoustic window, all middle cerebral artery occlusions were diagnosed with transcranial color Doppler sonography, as verified by angiography.
Blood flow velocity in a major cerebral artery can increase not only because of narrowing of its lumen but also because of hyperemia, caused either by a functional decrease in the peripheral vascular impedance or by a large arteriovenous malformation [31, 32]. One cannot differentiate spasm from hyperemia with conventional transcranial Doppler sonography, whereas transcranial color Doppler sonography allows direct visualization of flow within the artery in color; consequently, measurement of the vessel caliber may be attempted. The literature contains no information as to how accurate such measurements are, and our experience in this matter is limited. We were not able to identify any cases of hyperemia, and it is unclear whether this contributed to our false-positive results. Nevertheless, all 10 large arteriovenous malformations were directly visualized with transcranial color sonography.
False-negative findings can be explained by the influence of all the factors that reduce the flow velocity in the examined artery. Increases in intracranial pressure and other factors such as arteriosclerosis, disturbed autoregulation, thromboembolism, significant stenosis of carotid arteries, and decreased cardiac output, must also be taken into account [33].
In this study, all estimations of transcranial color Doppler sonography in diagnosing middle cerebral artery spasm were calculated for angiographic diagnoses for this condition. Digital subtraction angiography is often regarded as an excellent tool for diagnosing middle cerebral artery spasm, but it also has limitations [34]. The classification of arterial narrowing has inherent inaccuracies because vasospasm is sometimes not focal, and comparison with the adjacent segment does not necessarily provide a comparison with the unaffected artery. These problems are illustrated by our figure of interobserver agreement. Kappa statistics were less than perfect, although they did fall in the excellent range (0.8) for the diagnosis of middle cerebral artery spasm, and in a good range (0.63) for classification of middle cerebral artery narrowing. Repeated angiographic examination could help in the diagnosis but is rarely performed.
It would be worth comparing the accuracy of MR and CT angiography with transcranial color Doppler sonography in the definitive diagnosis of middle cerebral artery spasm. No study of the same population using these two modalities has been published, so only indirect comparison may be attempted. Grandin et al. [35] compared the accuracy of time-of-flight MR angiography with the accuracy of digital subtraction angiography. The specificity of MR angiography was found to be very good (99%), whereas the sensitivity was less (only 56%) than that found in our study. This rather poor sensitivity of MR angiography in detecting vasospasm of the middle cerebral artery can be explained by the often tortuous course of the vessel, artifacts related to patient movement, and the inhomogeneous flow velocity distribution in the artery as a result of increased velocity in the narrowed segment [36]. Additionally, the presence of an aneurysm clip or coils and degraded hemoglobin around the vessel can blur its outlines at the point where vasospasm is actually likely to develop.
CT angiography is potentially an ideal method for assessing patients with subarachnoid hemorrhage for vasospasm because usually these patients require an unenhanced CT study of the head, at which time CT angiography can be added. CT angiography is highly sensitive and specific for the detection of severe vasospasm of more than 50% of lumen reduction [37]. Nevertheless, for detection of mild and moderate spasm, CT is much less accurate [37].
From our study we can conclude that the performance of transcranial color Doppler sonography is high in the identification of patients with moderate to severe middle cerebral artery spasm and only slightly lower for middle cerebral artery spasm of all grades. Standardization of blood flow velocities with respect to age and sex further increases the performance of transcranial color Doppler sonography, especially in relation to diagnosing less pronounced middle cerebral artery spasm. The optimal diagnostic peak systolic velocity threshold for patients older than 60 years is 156 cm/sec; for the middle age span, 185 cm/sec; for young men, 192 cm/sec; and for young women, as high as 213 cm/sec.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Rigamonti, A. Ackery, and A. J. Baker Transcranial Doppler monitoring in subarachnoid hemorrhage: a critical tool in critical care: [Monitorage par Doppler transcranien lors d'une hemorragie sous-arachnoidienne : un outil indispensable aux soins intensifs] Can J Anesth, February 1, 2008; 55(2): 112 - 123. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Krejza, W. Rudzinski, M.A. Pawlak, M. Tomaszewski, R. Ichord, J. Kwiatkowski, D. Gor, and E.R. Melhem Angle-Corrected Imaging Transcranial Doppler Sonography versus Imaging and Nonimaging Transcranial Doppler Sonography in Children with Sickle Cell Disease AJNR Am. J. Neuroradiol., September 1, 2007; 28(8): 1613 - 1618. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Krejza, J. Kochanowicz, Z. Mariak, J. Lewko, and E. R. Melhem Middle Cerebral Artery Spasm after Subarachnoid Hemorrhage: Detection with Transcranial Color-coded Duplex US Radiology, August 1, 2005; 236(2): 621 - 629. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
